1. AutoMedia: Linking the Vehicle with Consumer Electronics and Services Jason Flinn University of Michigan

2. Problem: Vehicle is an island Each vehicle is an island unto itself data crosses the border generally with carried device device (and its data) must be consciously managed limited extra-vehicle networks are separate stovepipes Several lost opportunities passengers don’t have what they want context which could help organize is ignored few existing applications don’t coordinate many vehicular applications hard to create Jason Flinn

3. Vision: Bridge the islands Bridges to the “mainland” of home and office unified view of all of a passenger’s media automatic exchange of context and meta-data allow individual consumer devices to come and go no effort on the part of the passengers Bridges to the “other islands”—vehicles on the road exchange traffic, road conditions communicate for immediate safety concerns form communities among co-located travelers Jason Flinn

4. What’s needed? Advances in distributed storage: Share data among home, vehicle, and web services Cache data and propagate updates asynchronously Support non-traditional clients Advances in networking: Leverage multiple networks: WiFi, WiMax, GPRS, DSRC Predict where you are going Provide connectivity forecasts Advances in middleware: Allow applications to declare intentions Hide vagaries of intermittent and uneven connectivity Jason Flinn

5. What’s needed? Advances in distributed storage: Shares data among home, vehicle, and mobile devices Cache data and propagate updates asynchronously Support non-traditional clients Advances in networking: Leverage multiple networks: WiFi, WiMax, GPRS, DSRC Predict where you are going Provide connectivity forecasts Advances in middleware: Allow applications to declare intentions Hide vagaries of intermittent and uneven connectivity Jason Flinn

6. Managing Data: Today Jason Flinn

7. Managing Data: BlueFS BlueFS helps manage personal data and devices My Data My Devices Jason Flinn

8. BlueFS Goals Pervasive data access (anytime, anywhere, any device) Support consumer electronics devices (iPod, phone, etc.) Support web services (Flickr, Facebook, Amazon S3, etc.) Simplified data management Automation (indexing, organizing, transcoding, etc.) Policies (reliability, privacy, etc.) Context (location, time, nearby people, etc.) Jason Flinn

9. BlueFS Overview Server stores primary replica of all data Clients cache data to improve performance Clients register affinity for specific types of dataCached data available when disconnected Jason Flinn

10. Device-Specific Protocol Connecting Consumer Electronics CEDs cannot perform DFS protocol DFS Protocol? Distributed File System Device-Specific Protocol Attach: computer speaks for CEDWorks with disconnected clients DFS Protocol Jason Flinn

11. Device-Specific Namespaces User prefers one organization Consumer electronics device requires another organization Solution: Store translations on the consumer device Classical Jazz F01 F02 F03 F04 Jason Flinn

12. Adding Automation Updating indexes Transcoding Type-specific caching Organizers Jason Flinn

13. Adding Automation Cannot execute on CED Leverage general-purpose computers Take action when files change Problem: Need notification of file changes Jason Flinn

14. Notification Via Persistent Queries Don’t need new mechanism Leverage cache consistency mechanism Structure notifications as file system object Robust to crash Handles disconnected operation updatestruncate readprocess *.mp3 Jason Flinn

15. .TiVo.mp4 Persistent Queries Example: video recording and playback File Server Video Transcoder *.TiVo.mp4.TiVo *.TiVo.TiVo.mp4 Jason Flinn

16. Leveraging intra-vehicle context Vehicle has a wealth of contextual data Can we provide a “context bus” that enriches data? Enrich data by sharing context among local devices: vehicle location used to tag site where photos taken Allows searching by location Can automatically organize digital photo albums Wireless proximity of cell phones reveals user presence Match music on stereo to tastes of passengers Start playing audio-book where left off last time Jason Flinn

17. What’s needed? Advances in distributed storage: Shares data among home, vehicle, and mobile devices Cache data and propagate updates asynchronously Support non-traditional clients Advances in networking: Leverage multiple networks: WiFi, WiMax, GPRS, DSRC Predict where you are going Provide connectivity forecasts Advances in middleware: Allow applications to declare intentions Hide vagaries of intermittent and uneven connectivity Jason Flinn

18. The Big Picture State of the art: systems are reactive not proactive There are opportunities to plan people are creatures of habit infrastructure is slow to change If the system knew where you were likely to be, and what connectivity was available there......then applications could schedule their usage appropriately Our goal: expose this connectivity derivative to applications Jason Flinn

19. Pieces of the puzzle Where are you likely to be? figure out where you are now use the past to predict where you’ll be next What are conditions likely to be when you get there? map access points at newly-visited locations refresh infrequently to catch stale state Expose this information to applications simple API: expected bandwidth in 10 (20, 30,...) seconds applications use this however they like Jason Flinn

20. Where are you? Existing solutions GPS when you can WiFi localization GSM fingerprinting Our implementation: PlaceLab/WiFi 20-30 meter avg accuracy acceptable for our needs Convert to a discrete grid limit precision of lat/lon grid square ~110x80 meters compact, useful representation Jason Flinn

21. Where are you going? We need a personal mobility model people are creatures of habit, but......different people have different habits Song et. al. (Dartmouth) conducted a bake-off winner: second-order Markov model Model consists of discrete states and transitions states: your current and former location transitions: possible next states, with probabilities assigned probabilities computed by past behavior discrete transition steps: we use 10 seconds (arbitrary) Jason Flinn

22. What will conditions be like? Overlapping public coverage Heterogeneous quality Bandwidth, latency, ports Decentralized ownership Multiple “usable” options Which AP is the best AP? Jason Flinn

23. Virgil AP selection daemon Current mechanism: loudest must be best okay in a homogenous deployment no better than random selection in the wild Instead, we perform active testing and discovery Scan for APs and associate to each in turn Run bandwidth, latency, port tests to reference server Choose AP with the “best” connection Cache test results for future prediction Jason Flinn

24. History is useful Repeated traversals of the same area benefit (a lot!) History: % scans finding usable AP History: scanning overhead (s) ‏ Jason Flinn

25. The Commute Jason Flinn

26. Connectivity Forecasts Close is often good enough, thanks to range issues Jason Flinn

27. What’s needed? Advances in distributed storage: Share data among home, vehicle, and mobile devices Cache data and propagate updates asynchronously Support non-traditional clients Advances in networking: Leverage multiple networks: WiFi, WiMax, GPRS, DSRC Predict where you are going Provide connectivity forecasts Advances in middleware: Allow applications to declare intentions Hide vagaries of intermittent and uneven connectivity Jason Flinn

28. Declarative Networking Allow applications to declare their intent in using network. Main goal: simplicity of the interface. Applications specify: Foreground (latency-sensitive) vs. background Bulk vs. small messages Important vs. unimportant etc. May specify on a per-message, per-socket, or per-thread basis. Jason Flinn

29. Connectivity middleware Manage vagaries in network connectivity: Transient networks (WiFi hotspots) Intermittent networks (cell coverage) Changes in latency Varying bandwidth Challenge: map application sends/receives to networks Choose best network for each request Defer traffic when necessary, continue when available Allow applications to tear-down, re-establish connections Jason Flinn

30. Declarative networking example Read mail (foreground) High availability Low bandwidth Intermittent High bandwidth Download mail (background) Jason Flinn

31. Building Bridges Advances in distributed storage: Share data among home, vehicle, and mobile devices Cache data and propagate updates asynchronously Support non-traditional clients Advances in networking: Leverage multiple networks: WiFi, WiMax, GPRS, DSRC Predict where you are going Provide connectivity forecasts Advances in middleware: Allow applications to declare intentions Hide vagaries of intermittent and uneven connectivity Jason Flinn

32. Thanks to… T.J. Giuli Brett Higgins Anthony Nicholson Brian Noble Dan Peek Venkatesh Prasad Azarias Reda David Watson Jason Flinn